← 返回
未分类 中文

Maverick Process To Skill

Turn a user-described business process into an automated execution flow and optionally convert it into a reusable local skill. Use when a user explains steps...
将用户描述的业务流程转换为自动化执行流程,并可选择将其转换为可重用的本地技能。用于用户解释步骤时...
maverick maverick 来源
未分类 clawhub v1.0.0 1 版本 100000 Key: 无需
★ 0
Stars
📥 240
下载
💾 1
安装
1
版本
#latest

概述

Process To Skill

Use this skill to operationalize "explain once, reuse later" behavior for business workflows.

Workflow

  1. Extract the process from the user's message as an ordered list of steps, inputs, and expected output.
  2. Execute or automate the process for the current request.
  3. Confirm the result in plain business language.
  4. Ask whether the user wants to save the process as a reusable skill.
  5. If the user says yes, create a new skill folder with a trigger-focused description, concise workflow instructions, and agents/openai.yaml metadata.
  6. Return the saved skill name and one example command the user can use next time.

Save Prompt Requirement

After completing the process, always ask a direct reuse question:

  • "Do you want me to save this as a reusable skill so next time you can call it directly without re-explaining the steps?"

If the answer is unclear, ask once more with a suggested skill name.

Skill Authoring Rules

  • Use lowercase hyphenated names (- when possible).
  • Keep SKILL.md procedural and short; include only repeatable steps.
  • Put trigger conditions in frontmatter description, not in the body.
  • Add only required resource folders (scripts/, references/, assets/) when they are actually needed.
  • Reuse existing local scripts/templates when possible instead of duplicating logic.
  • Minimal Capture Template

    When saving a process, capture:

    • Goal: What business outcome this process produces.
    • Inputs: Required user-provided data.
    • Steps: Deterministic sequence to run.
    • Output: Expected deliverable format.
    • Trigger examples: 2-3 natural language prompts that should activate the skill.

    Guardrails

    • Do not auto-save without user confirmation.
    • Do not save one-off ad hoc tasks that are unlikely to be reused.
    • If required tools or permissions are missing, report the blocker and offer a partial skill draft.

    版本历史

    共 1 个版本

    • v1.0.0 当前
      2026-05-21 15:09 安全 安全

    安全检测

    腾讯云安全 (Keen)

    安全,无风险
    查看报告

    腾讯云安全 (Sanbu)

    安全,无风险
    查看报告

    🔗 相关推荐

    design-media

    Maverick Canva Mcp

    maverick
    通过Canva托管的MCP服务器搜索、读取和管理Canva设计作品。轻量级传递至Canva官方MCP;实时工具目录取决于...
    ★ 0 📥 816
    ai-agent

    Find Skills

    root
    帮助用户发现和安装智能体技能,当用户询问如「如何做X」、「找X的技能」、「有能做...的吗」等问题时
    ★ 1,523 📥 580,015
    ai-agent

    self-improving agent

    pskoett
    记录自身发现以实现自我改进的技能
    ★ 4,170 📥 942,377